Changes in Performance on Common Return-to-Sport Tests During a Collegiate Women’s Soccer Season: Implications for Baseline Testing

Abstract

Background:

Baseline testing of objective lower limb function may help clinicians make more informed return-to-sport (RTS) decisions in the event of an anterior cruciate ligament (ACL) injury. However, as these tests are based on physical performance, it is possible that they improve during the season as athletes get stronger and fitter. Hence, it may be difficult to ascertain the patient’s preinjury status and have an accurate reference for comparison when determining readiness for RTS. The purpose of this study was to examine changes in common ACL RTS tests during a college soccer season to determine the most appropriate time to perform baseline testing.

Hypothesis:

Hop test performance will improve across the season.

Study Design:

Descriptive laboratory; prospective cohort.

Level of Evidence:

Level 4.

Methods:

A total of 31 women’s soccer players from 1 NCAA Division I university agreed to participate this study. Participants performed 4 single-leg hop tests and the 505-agility test to measure cutting speed on 3 occasions: preseason (PRE), midseason (MID), and end-of-season (END). Performance on each test was compared across days to determine whether performance increased during the season. As a secondary analysis, limb symmetry index (LSI) was also compared across the season.

Results:

A total of 23 participants (age, 19.7 ± 1.3 years; height, 1.69 ± 0.07 m; weight 60.9 ± 7.2 kg) completed all 3 testing sessions during the season. Performance during PRE was better than MID and END for all hop tests (all P < 0.01). LSI did not change during the season (P value range, 0.18-0.79).

Conclusion:

Performance on the hop tests was best during preseason and declined thereafter, which may be reflective of cumulative fatigue.

Clinical Relevance:

Baseline testing of RTS tests performed during preseason may provide an accurate representation of an athlete’s best abilities over the course of a collegiate soccer season. Preseason testing would also enable athletic trainers to acquire baseline data for all athletes before they are injured.

Determining when a patient is ready to safely return to sport (RTS) after anterior cruciate ligament (ACL) injury remains a challenge for sports medicine practitioners across all settings. That up to 25% of adult athletes and one-third of youth athletes go on to suffer a second ACL injury after returning to sport,^24,30 particularly in the first 2 years after ACL reconstruction (ACLR),^30,31 indicates that athletes are cleared for unrestricted activity before they are properly rehabilitated and reconditioned. One of the contributing factors to this problem is likely the overall lack of objective functional criteria used in the RTS decision-making process. ³⁰

While the consensus is that no single particular criterion or test will suffice for determining readiness for RTS or identifying injury risk, hop tests may be the most clinically feasible way of mass screening or baseline testing for athletes compared with other measures such as isokinetic dynamometry or laboratory-based movement analysis. Given the minimal-to-moderate correlations between hop tests and isokinetic strength tests,^17,28 and that quadriceps weakness has been shown to underlie larger asymmetries in hop tests in ACLR patients,^29,33 persistent weakness post-ACLR or that may predispose one to future injury may be initially detected without the need for a dynamometer. Since asymmetry has been associated with injury and, specifically, triple hop distance, ¹⁸ limb symmetry index (LSI) has been included in a high-risk profile for second ACL injury, ³⁰ performing baseline testing may also be used as a screening tool to identify those with a deficit and in need of intervention.

When objective functional assessments are performed, a battery of single-leg hop tests is common because of their similarity with functional sport maneuvers and easy, low-cost administration. ⁷ When interpreting hop test performance, the performance of the ACLR limb during various tasks is typically compared with the uninjured limb, which serves as a “control” limb, and expressed as a LSI. Achieving an LSI of 90% is considered by most to be the standard for “passing” the test ¹ ; however, a limitation of using LSI in this way is that function may be overestimated since the uninvolved limb has likely become deconditioned during the lengthy rehabilitation process as well. In previous work, the ACLR and uninvolved limbs both performed below that of normative data from healthy controls despite having exceeded the 90% LSI benchmark, ¹⁵ which demonstrated that detraining of the uninvolved limb during rehabilitation contributed to an inflated LSI. To address that limitation, a study measured the capabilities of the uninvolved limb around 2 months postinjury as an estimate of their preinjury status and found that, of all patients that met the traditional 90% LSI criterion, that only about one-half of those patients achieved 90% of the estimated preinjury capacity. ³⁴ Collectively, these findings provide further evidence of the decrease in function of both limbs after surgery and that having preinjury data would be ideal rather than relying on LSI. ²²

While baseline testing for concussion has become the standard of care in sports medicine, few athletic trainers are performing baseline testing for lower extremity injury RTS criteria. However, unlike neurocognitive testing for concussion performed at the beginning of each year, RTS tests are physical-performance-based and are therefore subject to change during the season as the athletes assumingly get stronger and faster with training. Thus, preseason performance may not be an accurate indicator of their abilities later in the season. However, this is somewhat conjectural as other factors during the season, such as cumulative fatigue, injury, motivation, etc, may also contribute to a decline in performance. To our knowledge, no one has investigated changes in lower limb function tests during a competitive athletic season, but there is a small body of literature in soccer players reporting decreases in some performance tests during the season,^4,25 which were attributed to a net catabolic effect of the physiological load experienced during a competitive season. As such, we do not know the appropriate time of the season to collect baseline RTS test data to give us a metric of the athlete’s best ability to use as a benchmark should injury occur. Therefore, the purpose of this study was to compare performance on common RTS tests across a NCAA Division I Women’s Soccer team during the competition season. Based on equivocal evidence of changes in performance on various metrics and the absence of evidence that hop tests change over time, we hypothesized that performance on these tests would improve during the season as the athletes got stronger and fitter with training.

Methods

Design and Participants

We recruited the women’s soccer team at an NCAA Division I institution to participate in this prospective cohort study before the start of their Fall competitive season. We chose women’s soccer players because this population has consistently accounted for a disproportionate number of noncontact ACL injuries since injury surveillance efforts began in 1983.^2,3,5,14,19 To participate, they must have been medically cleared by the team physician and athletic trainer for unrestricted team training and competition. Exclusion criteria for any given testing day included having any pain or dysfunction that would interfere with their ability to perform the maximal effort tests. Participants provided their consent before participation in accordance with the university’s institutional review board procedures. A total of 31 players (age, 19.6 ± 1.3 years; height, 1.69 ± 0.07 m; weight, 62.1 ± 8.8 kg) met the inclusion criteria and agreed to participate. Upon enrollment in the study, we documented the participants’ dominant limb, defined as the leg they would prefer to kick a ball for maximum distance.^11,16,26

Procedures

The participants completed 3 identical testing sessions, separated by 6 weeks to divide the season equally: on the day before preseason training began (PRE), at midseason (MID), and during the last week of regular season competition (END). Each testing session occurred at the same time (7:00 am). The MID and END testing days were chosen purposefully to occur after a scheduled day off training and competition and both occurred after a typical weekend of conference play. Between testing sessions, all participants completed their regular team training regimen, consisting of soccer-specific training and strength and conditioning.

To increase external validity and the participant’s comfort and familiarity with the tasks, each testing session was conducted outside on the natural turf field, with participants in their self-selected soccer cleats. The sessions began with 2 minutes of jogging for general tissue warming and then a standardized 8-minute dynamic flexibility warm-up, ¹¹ followed by practice trials of all the tests. Testing was performed in a “mass screening” format to aid in efficiency and to represent the ideal format for baseline testing. Participants were randomly split up to different stations and rotated to the next station as a group once they were finished with the required trials. Participants were familiarized to all tests during their summer training program and again during practice trials as a part of the warm-up at the beginning of each testing session. Three successful trials of all tests were completed on each limb, in a random order determined by the researcher at the station who would call out the limb to be used in that trial. A successful trial was considered if the participant completed the hop or cutting tests without losing their balance, falling, or slipping. If any of these mistakes occurred, the participant redid the trial. Participants received at least 1 minute of rest between trials while they waited for others in their group to complete a trial.

RTS Tests

We selected the tests used most commonly to assess readiness for RTS, including single-leg hop tests (single, triple, crossover, 6-m timed).^10,27,32 While not a common RTS test, we included the 505-agility (505) test so we could get a metric of soccer-specific demands of full-speed sprinting and single-leg cutting ability. ²¹ For the single, triple, and crossover hop tests (Figure 1), standard tape measures were staked into the ground to measure hop distance (m). Participants were asked to stand with their toe on the zero mark of the tape, shift into a single-leg stance, pause, and then hop as far as possible and “stick” the landing. Their landing spot was then marked and recorded. For the triple hop, participants were instructed to perform 3 maximal consecutive hops without pausing. For the crossover hop test, 3 consecutive hops were performed while crossing over alternate sides of the measuring tape. For the timed 6-m hop (Figure 1), the participants were asked to assume the same starting position as the other hop tests. Then, on their mark, they hopped as fast as they could through the finish line placed 6 m away while infrared timing gates (TCi system; Brower Timing Systems) recorded their time (s).

Figure 1.

The 505-agility test.

The 505-agility test consists of a 15-m sprint, a 180° single-leg cut, and a 15-m return sprint. Infrared timing gates (TCi system; Brower Timing Systems) were placed at the starting line and at 10-m to record the 5-m cut time (Figure 1). Participants were instructed to sprint down and back as quickly as possible. Cut time (s) was measured as the time elapsed between the instant that the participant broke the plane of the timing gate placed 5 m away from the end line, made a single-leg cut at the end line, and returned through the same timing gate.

Statistical Analysis

The best of 3 trials (maximum distance [m] for the single, triple, and crossover hop tests and minimum time [s] for the 6-m timed hop and 505-agility tests) was used for the analysis to represent the participant’s maximum ability for each limb. ⁹ For the hop tests and the 505-agility test, LSIs were calculated by dividing the lower-performing limb by the better-performing limb and multiplying by 100, resulting in a value that expressed the performance of the lesser-performing limb as a percentage of the better-performing limb, which could be interpreted as a between-limbs “deficit.” All data were inspected for normality. The individual limb data were found to conform to the assumptions for parametric testing, but the LSI data did not follow the normal distribution.

To determine whether performance on each test changed over the course of the season, separate 2 (Limb) by 3 (Day) repeated measures analysis of variance (ANOVA) tests were performed. Post hoc t tests with Bonferroni correction were used in the event of a significant omnibus test. In addition, to determine whether the LSI changed over the course of the season, the Friedman test was performed as a nonparametric analogue to a repeated-measures ANOVA. Significance was set a priori at P < 0.05.

Results

A total of 31 participants completed the preseason testing session. Over the course of the season, 8 participants missed ≥1 of the remaining testing sessions due to injury or disqualifying pain. Thus, 23 participants (age, 19.7 ± 1.3 years; height, 1.69 ± 0.08 m; weight, 60.9 ± 7.2 kg) were included in the final analysis. Descriptive data for all tests are provided in Table 1. Full ANOVA and post hoc results are available in Appendix Tables A1 and A2, respectively (available in the online version of this article).

Table 1.

Descriptives (Mean [SD] for performance on each test during the PRE, MID, and END are provided for the dominant and nondominant limbs

Testing Period	Limb	Single Hop (m)	Triple Hop (m)	Crossover Hop (m)	6-m Timed Hop (s)	505-Agility Test
						5-m Cut (s)
PRE	Dominant	1.98 ± 0.19	5.46 ± 0.49	5.11 ± 0.76	1.90 ± 0.16	2.54 ± 0.12
	Nondominant	1.99 ± 0.15	5.47 ± 0.44	4.88 ± 0.47	1.86 ± 0.16	2.49 ± 0.11
	LSI a	95.0 ± 4.1 (83.6-99.6)	95.1 ± 3.6 (85.1-99.7)	92.9 ± 5.8 (78.0-99.6)	93.3 ± 3.8 (85.6-98.9)	97.1 ± 2.0 (93.9-99.6)
MID	Dominant	1.57 ± 0.16	4.69 ± 0.43	4.39 ± 0.38	2.03 ± 0.15	2.49 ± 0.16
	Nondominant	1.55 ± 0.17	4.82 ± 0.43	4.35 ± 0.41	1.96 ± 0.18	2.50 ± 0.22
	LSI a	93.6 ± 6.0 (75.0-100.0)	93.8 ± 5.2 (78.4-99.8)	93.6 ± 5.0 (82.5-99.8)	93.9 ± 5.0 (82.5-99.5)	95.6 ± 4.9 (78.7-100.0)
END	Dominant	1.57 ± 0.13	4.69 ± 0.40	4.22 ± 0.38	2.06 ± 0.17	2.53 ± 0.10
	Nondominant	1.56 ± 0.12	4.65 ± 0.36	4.14 ± 0.41	2.04 ± 0.17	2.53 ± 0.12
	LSI a	94.5 ± 4.4 (82.4-99.4)	94.4 ± 3.5 (87.1-99.2)	92.9.2 ± 4.6 (83.3-99.8)	92.6 ± 4.8 (81.9-100.0)	97.0 ± 2.8 (90.5-100.0)

^aLSI (Limb_min/Limb_max) × 100)) and LSI range (min-max) are provided.

END, end of season; LSI, limb symmetry index; MID, midseason; PRE, preseason.

Changes in RTS Test Performance During the Season

Significant differences in performance were noted for all assessments across days: (F_2,66 range, 8.7-71.3; all P < 0.001) except for cutting time during the 505-agility test (F_2,66 = 0.6;P = 0.58) (Figure 2). Post hoc tests revealed that participants hopped farther during PRE versus MID (all P < 0.001) and versus END (all P < 0.001) with no differences between MID and END (P value range, 0.38-1.00) on the single, triple, and crossover hop tests. For the 6-m timed hop test, tests participants hopped faster during PRE versus MID (P = 0.02) and versus END (P < 0.001), with no difference between MID and END (P = 0.54).

Figure 2.

Hop test performance across the season. *Pre > Mid, ^†Pre > End, ^§Pre < Mid, ^‡Pre < End.

Differences Between Dominant and Nondominant Limbs

No differences between the dominant and nondominant limbs were observed for any test (F_1,66 range, 0.31-3.96; P value range, 0.05-0.58; Cohen’s D effect size range, 0.04-0.37). There were no Limb by Day interactions for any test (F_2,66 range, 0.27-1.31;P value range, 0.28-0.76).

Changes in LSI Across the Season

No differences in LSI were observed for any test across the season (χ² range, 0.46-3.45; P value range, 0.18-0.79; Kendall’s W effect size range, 0.01-0.08).

Discussion

We compared performance on the tests used most commonly to assess readiness for RTS after ACLR during a collegiate soccer season. Participants performed best on the tests during PRE and their performance declined at MID and remained there at END. We hypothesized that performance would improve on the tests we examined in this study as we had no evidence that these particular tests were susceptible to declines over the course of the soccer season. There are several practical explanations for our findings, namely cumulative fatigue and decrements in physical condition.

Previous work has investigated changes in performance on various physical performance tests during a soccer season. While contrary to our hypotheses, our findings are consistent with previous studies of collegiate soccer players from the strength and conditioning literature that reported reductions in isokinetic knee extensor strength, vertical jump height, and sprint speed over the course of the season in men, ²⁰ and reductions in knee extension isometric peak torque and rate of force development in women. ⁴ Those authors attributed these decrements in performance to the catabolic effects of training and competition demands as the season wore on, despite their conditioning program. Collectively, these findings hold relevance for coaches as it may point to a need to more closely examine their training programs to allow for adequate recovery to achieve the peak performance that is desired at the end of the season. Instead, our data show that the athletes were experiencing a midseason slump in performance. The decline in hop test performance may also be reflective of lessened neuromuscular control, which could represent an increased risk of injury.

We were unaware of any existing literature that investigated changes in hop test performance over the course of the season. The current study adds to the relatively new body of literature some additional evidence that maximal physical performance degrades over the course of a soccer season, likely due to the development of cumulative fatigue. We also noted that the participants in the current study exceeded the normative values proposed for collegiate women’s soccer players in previous work, ²⁷ which highlights a limitation of using normative data, although it has been proposed to be a stronger reference than the uninjured limb. ²²

Practically speaking, we also attribute the decline in performance to what could commonly be called “aches and pains” that developed as the season wore on and that are commonly observed. While participants did not participate in testing on any given day if they were injured as defined in our exclusion criteria, anecdotally, the athletes complained of more “soreness,” “tightness,” etc, and sought athletic training services more frequently for treatment (ie, the athletic training clinic was visibly full of soccer players as the season progressed), without any diagnosed pathology. This observation is probably reflected in recent work that reported a continuous reduction in vigor in women’s soccer players (as measured by the Multi-component Training Distress Scale) over the course of the season,^23,25 as well as an increase in depression at season’s end. However, this remains somewhat conjectural since we did not administer a patient-reported outcome instrument to measure their impairments and health-related quality of life. It remains plausible that, collectively, these factors could lessen an athlete’s motivation to complete maximal effort testing as the season progresses despite being “healthy.”

We observed a decline in performance outcomes (ie, distance hopped and time to completion) after the preseason period, but no changes in the LSI, which indicates that both limbs changed similarly. However, it should be noted that our results are based on the group averages. When looking at individual-level data, we did see variability in both the magnitude of change in test performance during the season and LSI. That is, some participants’ performance and/or LSI did not change during the season while others did. The variability across participants is similar to those noted in previous work that tracked limb symmetry during countermovement and drop jumps across a soccer season to examine their associations with other performance outcomes. ⁶ This observation is certainly interesting and deserving of further investigation; however, it would be very difficult without documentation of other factors that could contribute to this variability, such as playing time, position, injury history, etc, which we did not have.

The rationale for this study was to determine the best time to do baseline testing so that the data could be used to assess progress during rehabilitation and ultimate readiness for RTS should an ACL injury occur. But importantly, the data acquired from baseline testing could also be used to potentially identify those who may be at risk of injury. Although researchers have not been particularly successful at prospectively identifying those that will go on to suffer an ACL injury via clinical screening tests, assessing symmetry in “healthy” athletes might be particularly relevant in the soccer population as a recent meta-analysis suggests that the dominant (kicking) limb is more likely to be injured in soccer. ¹³ While many think of competitive soccer players being equally good with both feet, the literature indicates a high level of footedness at the highest levels of soccer competition. ⁸ Soccer kicking places asymmetrical demands on the lower extremity, with the nondominant (ie, “plant”) limb being responsible for more deceleration and stabilization while the dominant limb kicks the ball, it would be reasonable to expect some asymmetry. While the literature investigating between-limb symmetry in lower extremity strength and range of motion in small cohorts of soccer players remains equivocal, a recent systematic review and meta-analysis found that soccer players across age, sex, and competition level were symmetrical in strength, ¹² despite evidence of typical asymmetry in other soccer-specific functional measures. ¹² This evidence suggests that observing asymmetry during baseline testing with hop tests to provide a basic representation of strength could be considered abnormal and may merit additional attention. However, greater understanding of the range of asymmetry which could be considered “normal” is needed.

The LSIs were, on average, 94% across all tests in the current study, which is consistent with normative data reported from healthy, physically active adults.^15,22 However, we did observe clinically significant differences between limbs in some participants, as evidenced by the range of values for LSI (Table 1). Across all 3 days and hop tests, we observed 53 instances of LSI dropping below the 90% threshold in this cohort of players who reported being free from pain and injury on each testing day. ¹ Although limb symmetry is typically calculated for the involved versus uninvolved limb, in this study we calculated it as the lower-performing limb relative to the better-performing limb since none of the players had an injured limb at the time of testing. In that case, we were able to see an absolute deficit, regardless of limb. Most incidences of asymmetry occurred in the crossover and 6-m timed hop tests, which are the most dynamic and incorporate the most balance, which may point to more of a coordination issue than a strength issue. This is supported by recent work that reported that the LSI during hop tests was higher than the LSI from isokinetic strength testing, ²⁸ furthering the assertion that hop tests are not perfect surrogates for strength testing. But, interestingly, there were no instances of asymmetries below 90% in the 5-m cut time, which consists of a full-speed sprint to a single-leg cut. It is reasonable to surmise that soccer players more often perform cutting/change of direction bilaterally than kicking and would, therefore, be more skillful and symmetrical. As the sprint-cut maneuver is certainly performed more frequently than maximal single-leg hopping, that could help explain why this was the only test that did not worsen during the season.

Practical Considerations for Baseline Testing

The rationale for conducting this study was to address the practicality of performing baseline assessments of tests commonly used to help make RTS decisions after lower extremity injury, and particularly ACL injury. While our analysis determined that the preseason testing best represented the athletes’ maximal abilities and would therefore be the goal to which the athlete would ideally aspire during the rehabilitation process, this study left us with several practical considerations that further support the collection of baseline data during preseason.

Over the course of the study (ie, the competitive season), 8 players were unable to complete the testing due to injury on either the MID or END testing days or both. In that case, if baseline testing were to be done at any other point after preseason, we would most likely miss the opportunity to test some players. Consequently, we would either be missing these data to inform our rehabilitation decisions should the player go on to be injured or we would have to test them individually once they were healthy, which poses a large burden on athletic trainers to follow up with multiple athletes.

Measuring baseline data at the beginning of the season can also be done efficiently and with several teams when done in a “mass screening” format. We collected our data on the first day that the team reported for preseason. We set up several stations with athletic training students and graduate students administering the various tests at each station. We continued to get more efficient at testing as the season wore on, with the last testing session being completed in just over 1 hour. The increased efficiency was due to factors such as setting up more stations so the lines were shorter, using scribes to write down the data as another research assistant took the measurement, etc. As with other mass screening or group interventions, being efficient appears to be the key in the willingness of coaches and athletes to participate. Additionally, once the personnel and stations are set up, this coordinated structure would allow for more teams to be tested in a matter of hours.

Limitations

The findings of this small study are limited to 1 women’s NCAA Division I soccer team and therefore are not generalizable. While the training and competitions are fairly uniform within a sport and division due to NCAA regulations, it is unknown how our findings could be generalized to men, other sports, or different levels of competition. In addition, the specific training program of a team could influence the results (eg, volume and intensity of strength and conditioning).

We analyzed the entire team’s longitudinal results; hence, we were unable to see whether some individuals’ performance changed differently (eg, some performance increased, others remained steady, etc), which is plausible considering recent work that showed highly variable symmetry scores across 18 male soccer players ⁶ ; however, our purpose was not to monitor or explain changes in performance during the season, but to look at the practicality of team-based testing and the best timing to do so.

We did not include a control group in our study. A true control group would require the same testing in a group of individuals who did not experience the factors that we thought might influence the changes in performance over the course of a competitive soccer season. That is, we would have had to compare the changes in these highly dynamic tasks to a group of nontraining women, which we thought provided little ecological validity. Thus, we determined that there would be limited value in that comparison and chose the longitudinal cohort design.

Conclusion

Athletes from 1 NCAA Division I women’s soccer team performed best on tests used commonly to assess readiness for RTS after ACLR immediately before their season started, compared with later times in the season. Preseason baseline testing can provide a reasonable representation of an athlete’s best physical capabilities during the same season and could eliminate clinicians’ reliance on postinjury LSIs. In the event that an injury occurs during the season, the baseline data can provide athletic trainers with objective data to guide the rehabilitation process and make more informed decisions regarding RTS with the ultimate goal of reducing the risk of future injury.